Does martingale really works? - page 69.

el kaka22

legendary

Activity: 3654

Merit: 1165

www.Crypto.Games: Multiple coins, multiple games

still, no matter how many coin that you have when using the martingale i'm sure will be lost all,, why? it is because we would eat greed want to win big in a short time ..
the worst that i have experienced while playing dice with a 50% chance to win is 23X is lose in a row.. how do you handle 23X lose in a row?
yaaa of course with a little basebet or just 1 satoshi it maybe safe but you're not going to settle for just a little win bitcoin,
then you will raise your basebet size, but is it possible for your balance can handle that lost streak?
sometimes if you are lucky you can win big with the martingale but not rarely you will be disappointed Cheesy

ranlo

legendary

Activity: 1988

Merit: 1007

Quote from: Hfleer on March 11, 2015, 08:55:33 PM

It is actually pretty easy to double your starting bankroll if you use 2x. No matter with what amount you start, the chance to double up will be aroung 50% - the houseedge you are betting with.

On the other hand. If you have unlimited funds, its impossible to double them. because you can't double unlimited founds per definition Wink

Let's say you start with 1 BTC, and you're going at 2x. It will depend on how much you're risking. Let's say you want to risk:

1 BTC at 0.00001 BTC starting (2x on loss) = you can lose 13x in a row. it will take 100000 wins before getting to doubled up (which is around 200000 rolls). Essentially you're dealing with (tl;dr version):

Won amount per win: 0.001%
Chance to lose 13x in a row: 0.00179%

You're statistically going to bust 1.8x for every 1x you double up, giving a 35% or so chance of making it happen.

Let's say you decide to risk 0.001 BTC per roll instead. Now you can lose up to 8x in a row. This puts the stats at:

Won amount per win: 0.1%
Chance to lose 8x in a row: 0.2136%

Now you're statistically over 2x more likely to bust than to double up, leading to a 30% or so chance of success.

If we did just one satoshi, you're at a 1.93:1 odd of busting before doubling up (with a 25x fail streak required).

Edit: someone check this and make sure I'm calculating the odds right? This is how I inferred how you'd figure out the true odds. I may be wrong on that...?

Hfleer

sr. member

Activity: 448

Merit: 250

Changing avatars is currently not possible.

Quote from: ranlo on March 11, 2015, 08:52:45 PM

Quote from: Hfleer on March 11, 2015, 08:38:30 PM

Quote from: ranlo on March 11, 2015, 08:34:43 PM

Quote from: Hfleer on March 11, 2015, 07:48:41 PM

Quote from: ranlo on March 11, 2015, 07:44:36 PM

Quote from: Hfleer on March 11, 2015, 07:41:11 PM

Quote from: JimClone on March 11, 2015, 04:18:45 PM

If martingale worked there would be a lot of dead dice sites but sadly I have tried to use it and end up losing all.

That is actually a good argument. People are still believing in martingale, so you could use that to convince them Wink

People believe in it because in theory, it would work. The reality, though, is that there are two things stopping it:

1) Player bankroll
2) Site max bet allowance

One of these is going to be hit, as neither of these can be infinite. Should these be removed, it would change things, as you could lose millions of times in a row and still not have "lost" yet.

Even in theory there is a third limitation: House could just stop anytime they want to. So if you are down 40 bets in a row, they could just say: "Thanks, good bye"

This is true as well, but we make decisions based on the assumption this won't happen. Otherwise we wouldn't be playing on the site/in that casino regardless as to if we wanted to martingale or not,

.

But it is a good point, that there's also the risk of them running. And in that case, when dealing with cryptos we can throw out another: the risk that the site is hacked (or "hacked") while you're trying to martingale.

That is the problem, there are so many limitations that even theoretical its impossible. But its a nice thought experiment nontheless. People have been grinding their brain on this for centuries for a reason.

The theory would state that you're dealing with a reputable (read: they won't run off) casino and the site/casino can't be hacked, though (or at least would be insured). So it really just deals with the numbers of betting vs. return.

I've run many simulations, trying to find a "sweet spot," including bets that differ from the 2x/50%, such as going to 3x, 4x, etc., and all end up the same way. With that said, I have found methods that statistically are a LOT safer than the straight 2x/50% way, but they all ultimately result in losses at some point.

The question has then always been to me... when do you stop? Mathematically speaking, you have a great chance to build up to SOMETHING. It's the continuing on that causes losses. I have yet to get anything built 2x its starting amount, for what it's worth, though I did come close with one (and it failed after repetition like 41k or so).